The common refrain among global-warming deniers is that mathematical models cannot make reliable predictions about something as complicated as the global climate. There is much value in that stance, and skepticism about model-generated predictions is not a bad idea in general. The climate of the planet is indeed extremely complex and many factors play a role. No climate scientist can really know with certainty how much weight various factors have at any given moment.
Certainty, however, is the key word. Deniers hold out the unreasonable standard of "certainty" when attacking the assertions of the Global Warming research. All these deniers do is inject a speck of doubt or uncertainty into the debate about climate change and then they strut about as though they've established a winning argument (this is the classic "muddy the waters" tactic). Well, we don't have to rely on models or episodic anecdotes to show how bad things have become regarding the state of the planet. See below.
A recent article in Sciencemag entitled "The Incredible Shrinking Sea Ice" tells the story of climatologists who wanted to check the models' predictions against what was happening in real time in the Arctic. Using aircraft and ship reports and satellite measurements they found that the IPCC's 18 climate models had UNDERestimated the rate of destruction of the Artic ice cap. The raw data from direct observation now suggests that Artic summers will be ice-free 30 years sooner than the models predicted, say between 2020 and 2070.
So the deniers are, in fact, correct. The models are wrong. They are far too generous in their estimations about how much time is left to correct the course we're on.
To repeat the line of last resort among the Bushies, "Who could have anticipated?!" Yeah, who could have seen this one coming? Only everybody not part of the military-industrial-entertainment complex.